CAR T cell therapy has been shown to be effective at inducing complete responses for acute lymphoblastic leukemia and other B-cell-related malignancies and has been shown to be effective at achieving and sustaining remissions for refractory/relapsed acute lymphoblastic leukemia (Maude et al., NEJM, 371:1507, 2014). However, dangerous side effects related to cytokine release syndrome (CRS), tumor lysis syndrome (TLS), B-cell aplasia and on-tumor, off-target toxicities have been seen in some patients.
There are currently two extant strategies to control CAR technology. The first is an inducible “kill switch.” In this approach, one or more “suicide” genes that initiate apoptotic pathways are incorporated into the CAR construct (Budde et al. PLoS1, 2013 doi:10.1371/journal.pone.0082742). Activation of these suicide genes is initiated by the addition of AP1903 (also known as rimiducid), a lipid-permeable tachrolimus analog that initiates homodimerization of the human protein FKBP12 (Fv), to which the apoptosis-inducing proteins are translationally fused. In the ideal scenario, these kill switches endeavor to sacrifice the long-term surveillance benefit of CAR technology to safeguard against toxicity. However, in vivo, these suicide switches are not likely to realize this goal, as they are operating against powerful selection pressures for CAR T-cells that do not respond to AP1903, a situation worsened by the inimical error-prone retroviral copying associated with the insertion of stable transgenes into patient T-cells. In this scenario, non-responsive CAR T-cell clones will continue to proliferate and kill target cells in an antigen-dependent manner. Thus, kill switch technology is unlikely to provide an adequate safeguard against toxicity.
The second CAR regulatory approach is transient CAR expression, which can be achieved in several ways. In one approach, T-cells are harvested from unrelated donors, the HLA genes are deleted by genome-editing technology and CAR-encoding transgenes are inserted into the genome of these cells. Upon adoptive transfer, these CAR T-cells will be recognized by the recipient's immune system as being foreign and destroyed, thus the CAR exposure in this system is transient. In another transient CAR exposure approach, mRNA of a CAR-encoding gene is introduced into harvested patient T-cells (Beatty, G L 2014. Cancer Immunology Research 2 (2): 112-20. doi:10.1158/2326-6066.CIR-13-0170). As mRNA has a short half-life and is not replicated in the cell or stably maintained, there is no permanent alteration of the CAR-expressing T-cell, thus the CAR expression and activity will be for a short period of time. However, as with the kill-switch approach, these transient CAR exposure approaches sacrifice the surveillance benefit of CARs. Additionally, with these transient systems acute toxicity can be difficult to control.